An Object-oiented Framework with Multi-objective Cellular Evolutionary Algorithm for Service Restoration of Shipboard Power Networks

Abstract

Pareto-based approaches clearly outperform conventional multi-objective approaches (i.e., the approaches of converting multi-objective into single objective by weighting factors), particularly when the preference relation among objectives often changes. However, when there are more than three objectives, for most Pareto-based approaches, it is very difficult to meet the speed requirement of service restoration of a shipboard power network. Due to lack of flexibility and compatibility, a fair comparison among algorithms is not easily implemented in process-oriented approaches. In this article, an object-oriented and multi-objective cellular evolutionary algorithm-based approach is proposed. In this technique, the multi-objective nature of the service restoration problem is retained, and the local search technology and elitist strategy are utilized. The proposed method greatly reduces the required execution time for the service restoration, and it has an obvious advantage, compared against other Pareto-based methods, concerning the diversity of solutions along the Pareto front. A public base platform has been constructed by extracting common attributes and operations of various Pareto-based methods. The computational procedures of three performance metrics are improved. As a result, a fair comparison can be achieved in the constructed platform. Simulation results demonstrate the feasibility and effectiveness of the proposed approach.